DeepSeek-V4-Flash-Vision-Uncensored Abliterated (refusal-removed) DeepSeek-V4-Flash for Ollama — ~305B MoE · native MXFP4 experts · thinking · 1M context

MIT · 1 tag · 156.4 GB · MXFP4 experts, bit-exact · text only · thinking

Website · Model Catalog · Hugging Face · GitHub · Discord · X

🛡️ New from Orca — OrcaCyber Zero 1.5

Post-trained for cybersecurity: vulnerability analysis, vulnerability reproduction, red teaming & offensive-security research. 98.01% pass@1 on CyberGym Level 1 · 1M context · function calling · OpenAI-compatible API.

Try it on OrcaRouter →

🔌 New from Orca — Orca Connect

Running this in Ollama on your own machine? Orca Connect puts the rig under your desk behind your usual OrcaRouter API key, next to 200+ hosted models, reachable from a laptop, a phone or a cloud agent. Outbound only: no port forwarding, no tunnel, no firewall rule. Requests your own hardware serves cost $0, and orca-connect claude points Claude Code at your rig for one session.

Connect your rig → · ▶ Watch the demo

DeepSeek-V4-Flash, abliterated, bit for bit.

The Ollama build of the GGUF conversion of orcarouter/DeepSeek-V4-Flash-Vision-Uncensored — the abliterated (refusal-removed) build of DeepSeek’s DeepSeek-V4-Flash-Vision-Exp, a ~305B / ~18B-active Mixture-of-Experts model with DeepSeek sparse attention, 4-wide Manifold-Constrained Hyper-Connections, a DSpark / MTP speculative head and a 1M-token context.

This is a format conversion, not a quantization. The 11,776 routed-expert matrices keep the base checkpoint’s own FP4 codes — DeepSeek’s per-32 ue8m0-scaled e2m1 layout is exactly GGUF’s MXFP4 block format, so they are repacked bit-for-bit, not requantized. There is no Q2_K…Q6_K ladder here and no quality/size trade-off to pick from: one tag, the same weights the parent repo ships.

⚠️ Disclaimer — read before use

This model has had its safety alignment substantially removed via abliteration (orthogonalizing the refusal direction out of the residual stream). It will comply with harmful, unethical, or illegal requests that the original DeepSeek-V4-Flash-Vision-Exp would refuse. Released strictly for legitimate research — interpretability, AI-safety / refusal-mechanism study, red-teaming, and robustness evaluation. You assume full responsibility for how you use it and everything it generates; add your own safety and moderation layers before any deployment. Use must comply with the MIT License inherited from the base model and all applicable law. The authors accept no liability for misuse.

⚠️ Text only — vision is not included

The base checkpoint is a vision-language model. These GGUFs, and so the Ollama build, are text-only. llama.cpp has no graph implementation for DeepSeek-V4’s vision tower — the only DeepSeek mmproj converters in the tree target DeepSeek-OCR, which expects a vision_config block and a SAM+CLIP tower that this checkpoint does not have. Rather than ship a broken projector, the converter drops the vision path entirely:

llama-cli reports modalities : text on load. Vision will be added here once llama.cpp supports the tower upstream — the vision residual writers are abliterated in the parent weights, so nothing needs re-baking, only re-exporting. For image input today, use the parent repo with DeepSeek’s own reference implementation.

Pick a tag

There is one tag because there is nothing to trade off.

Precision map: routed experts MXFP4 (bit-exact repack of the source FP4); attention wo_b, shared-expert w2 and the other block-FP8 matrices dequantized to BF16; Hyper-Connection mixers F32; hash-routing tables I32.

Usage (Ollama)

Requires Ollama v0.40.16 or later.

ollama run orcarouter/DeepSeek-V4-Flash-Vision-Uncensored

cURL

curl http://localhost:11434/api/chat -d '{
  "model": "orcarouter/DeepSeek-V4-Flash-Vision-Uncensored",
  "messages": [{"role": "user", "content": "Hello!"}]
}'

Python

from ollama import chat
resp = chat(
    model="orcarouter/DeepSeek-V4-Flash-Vision-Uncensored",
    messages=[{"role": "user", "content": "Hello!"}],
)
print(resp["message"]["content"])

• Thinking is on by default and comes back in message.thinking. Pass "think": false in the API, or --think=false on the command line, for direct answers.
• Give it room. With thinking on, set num_predict to 2048 or more so the answer is not cut off inside the reasoning.
• Prompting. Ollama renders prompts with the DeepSeek chat template embedded in the GGUF.

Evaluation

The abliteration was measured on the parent checkpoint, not separately re-measured on these GGUFs — see the parent model card for the harmful-refusal, over-refusal and capability tables. This conversion is a deterministic derivation of those weights: the expert matrices are bit-identical and the block-FP8 matrices are dequantized to BF16 without loss, so the behaviour carries over. Text generation was verified working on this build (llama-cli, 8×H100, ~57 tok/s), and the Ollama build passed chat tests with and without thinking.

Hardware

Only ~6 of 256 routed experts are active per token, so decode is far faster than 305B suggests, but the full weights must fit in RAM/VRAM (or be memory-mapped): budget ~146 GiB + KV cache. Ollama splits the model across GPUs and system RAM on its own.

License

MIT, inherited from deepseek-ai/DeepSeek-V4-Flash-Vision-Exp. Abliteration and format conversion do not change the underlying license obligations.

Elsewhere

• GGUF files: orcarouter/DeepSeek-V4-Flash-Vision-Uncensored-GGUF
• Full-precision source (safetensors): orcarouter/DeepSeek-V4-Flash-Vision-Uncensored

One Gateway. Every Model.

Route Smarter · Ship Safer · Spend Less

Website · Models · GitHub · Discord · X